The AM Forum

I recently went on a parts expedition down to Toronto, and snagged some 6146s! So I've started building something.

My homebrew power supply gives 300 and 500V out (it has other outputs too but I'm only using these two for this Xmitr).

I don't have much for modulation purposes...currently I'm using a pair of 6SK7s feeding a single 6L6 running from the 300V line. 6L6 cathode current is about 40mA (about 19 volts across the 470 Ohm Cathode resistor). Plate load for the 6L6 is a 30H 50mA choke. I have no audio Xformers and there's no place I can get any locally.

Plate B+ on the 6146 is 500V, in case anyone needs to know that. To monitor plate current, I have a 24V 5W lamp in series with the plate feed, prior to the plate RFC. From that I can monitor plate dip.

The plate of the 6L6 is connected to the screen grid of the 6146. But not directly connected because that would put 300V on the screen of the 6146! Instead I have a 10k resistor to drop the screen to about 170V and I have that resistor shunted with a 10uF 350V cap.

Right at the 6146's socket I have 3 X 150pF grounding caps to the screen and an RF choke in series with the DC feed to the screen.

When I'm trying to make a contact and I'm modulating (have my scope connected to the Xmitr's output to monitor the RF feeding the antenna) the modulation is really affecting only on the negative peaks. I'm not seeing positive peaks happening in the carrier.

It really sounds awful on my receiver (but I'm wondering if I'm overloading the receiver's front end merely due to close proximity).

When I scope the screen grid voltage on the 6146 I'm seeing my audio up to 150Vp-p so the audio isn't being cropped at the plate of the 6L6.

I previously tried using a Cathode follower (I have a few 6AS7 dual triodes) to drive the screen of the 6146 directly and had the same results in modulation - positive peaks didn't happen, only negative peaks came about.

I've tried running the 6146 at lower screen voltages (50V, 80V, 100V, 150V) - but the same modulation is observed on the carrier - negative peaks rule.

Has anyone ever seen anything like this?

Thanks and 73s
geo

My experience with screen modulation is that you have to increase the transmitter loading control past the point of maximum power output to get positive modulation.   That is something to try anyway.

I never did screen modulation, but besides being very inefficient, it seems like an interesting way of getting some hifi audio onto a carrier.

I think the screen has to be adjusted so the rf output is 1/4 full power with no audio, so when you modulate the screen, power can go up and down.

If you have a pair of 6146 tubes, they might do 100 watts of carrier, but in screen modulation, you might only get 25 watts, since 100% modulation would be 100 watts pep.

Unlike plate modulation where you ADD modulator power to the carrier, screen modulation takes power away at no audio but allows 100% power to pass on full modulation.

Most screen modulated rigs I have heard of were very low power.

Brett

I did some research and I've tried a variation on "Controlled Carrier" modulation:

http://www.noding.com/la8ak/f97.htm

I don't have the 6DE7 shown in the schematic, I do have other dual Triodes so I adapted the circuit to fit what I've got.

The circuit as shown modulates quite well but with a lot of distortion with the tubes I have.

K3ZS's advice was good, trying to keep loading up, that helped with the waveforms.
I have just a single 6146 running at 500V being driven by a 6AG7 oscillator.

Not sure if the 6146 is the best choice, maybe I need to switch to a 6BG6 or something?

The antenna is an inverted V cut for 1/2 wave 40m, hung from the top of the TV tower approx. 25 ft. maximum height, the ends tied to two small trees. My friend VE3AWA set it up and with his Kenwood he made quick contacts down to Indiana and West Virginia with this antenna.

I haven't made any contacts yet, but I'll keep trying.

73s
geo

GZB George,

I have never had much success with Screen Modulation but I do know that you need to drive the screen with a low impedance. Do you have a small step down transformer like a 5K to 600 Ohm military output transformer?

You also have to drive the screen negative by about 20 volts to get full modulation. Having the cathode well off ground can cause issues with the filament so be prepared to use a small isolated 6V transformer just for the 6L6.

The ARRL handbook always treated screen modulation very simply - it is not so simple to get clean distortion-free AM with this approach.

There was an interesting schematic in the old Editors and Engineers Handbook that claimed to solve the problems of non-linearity by using a cathode follower biased negatively at the cathode so it can swing below ground. By luck it uses a 6L6 and it would work on the 6146. I adapted the schematic and tied it to a representative (overly simple) RF section to show you how to hook it up.

Here is how they said to adjust it:

The loading of the RF stage should initially be adjusted normally. You will be adjusting the input power to the 6146 is 50% greater than the plate dissipation of the tube when we are done. The thing to remember is that if the carrier level screen voltage is correct for linear modulation of the stage, the loading will have to be somewhat greater than that amount of loading which gives maximum output from the stage.

R1 is Carrier level. R2 is Negative Peak Level. Now we have some controls to perfect the modulation envelope.

Initially set R2 for -50VDC at the pot wiper.

R1 should be adjusted until the carrier level screen voltage on the 6146 is about 1/2 of the rated screen voltage specified for Class-C CW operation. The current in the screen under carrier conditions will be about 1/4 of the normal current for CW conditions. You should try putting your choke in the grid to replace the 220K resistor at some point to see if that works better.

R2 should be adjusted so that negative modulation peaks, even when voltage which can cause overmodulation occurs, does not cause carrier cutoff.

The antenna coupling should now be increased while the carrier level is backed down by R1 until you get a distortion free modulated waveform in the scope.

Mike WU2D

The schematic is rather similar to what I've constructed this past week (but I didn't have the privilege of the Editors and Engineers Handbook).

Thanks for the input, I'll keep at it (while I'm still waiting on the 811s for the big transmitter).

73s
geo
VE3GZB

Sorry, no Xformers, but I like the cathode modulator - it's a nice low impedance solution.

If you notice the controlled carrier also uses a cathode modulator at the output.

Thanks and 73s
geo

Geo and others, at el....

I am trying to understand how my DX-60 controlled carrier works.  Mine has the WC3K audio mods too.

See http://www.amwindow.org/tech/htm/wc3kmods.htm (http://www.amwindow.org/tech/htm/wc3kmods.htm)

Geo,  the resting carrier control in that mod really does help a lot, especially if you ever use it to drive an amplifier.  In general it helps to tailor the modulation quite a bit.  Looking at W2UD's dual control is very intriguing.  Maybe it could be implemented in the DX-60.  The WC3K mod includes "clamping circuit" that flashes an LED (now installed on my front panel) when you overmodulate.  I can tell you it is VERY accurate.  Bright red flashes occur with every flat lining of the envelope on the scope.  

I had a qso with WB4BFS one morning about converting the DX-60 to pure SGM.  John are you following this thread?  I'd love for you to refresh my memory on what it would take to do that.  Is the RC in parallel on the cathode responsible for the controlled carrier?  It couldn't be as simple as removing them could it?

John KX5JT

You cannot completely screen modulate 100% without distortion.  The modulation linearity of the class-C stage goes south somewhere about 85-90% in the negative direction.  It is possible to maintain good linearity on positive peaks that extend above 100%.

I'm a little confused Don.  Are you saying that if the positive peaks extend above 100% "good linearity" can be maintained?  I understand that controlled carrier is intrinsically non-linear, since the carrier is changing in the output how can it be?  Omitting the actual carrier, what about the "audio" wave riding this changing carrier... is it able to be fairly linear in a controlled carrier enviroment... a pure screen grid modulated enviroment? 

I was perusing the links Derb posted in Terry's post on that Continental Broadcast site.  They used screen grid modulation. 

To quote from near the bottom of this page http://www.rossrevenge.co.uk/tx/continental.htm (http://www.rossrevenge.co.uk/tx/continental.htm)

"An even later version of the 316, the 316F is generally regarded as being the finest sounding AM transmitter ever built, with exceptionally low distortion and noise levels and excellent phase characteristics."

That seems pretty exciting to me!

Moderate modifications to a DX-60 produced an almost HIFi signal. Feedback was the key.
http://home.comcast.net/~suptjud/HomemadeHiFi.htm#Heathkit%20DX-60%20Audio%20Mods  (http://home.comcast.net/~suptjud/HomemadeHiFi.htm#Heathkit%20DX-60%20Audio%20Mods)

I agree - screen grid is not as linear as high level plate; the screen grid is square law in nature. But it can be improved to almost perfection as they did in the broadcast rigs.

First let him get it at least as good as a DX-60. Once you have it good...then we can play.

RF Feedback from a linear AM detector sampling the output to a remote cutoff driver stage like a 6AG7 control grid...

Mike WU2D

Yes, when I was working out what the resting screen grid drive should be, I found this to be true as well.

When arranging my circuit I saw I could leave the screen grid at a very low level and make the plate current swing nicely but then endure tons of negative clipping (bad). Or I could run the screen grid at a moderate level (between 100-200V) and get nice symmetry but not much plate current swing.

I chose the latter. Screen grid runs around 150V now and I'm using a scope to monitor my output so I don't overmodulate. If I get the plate loading just right then the output looks very traditional and proper.

Still, I've been trying to make contacts with it last night, and again this morning on 40m. Nothing.

I know the antenna is good (it seemed to be naturally resonant close to 7.1, that's where the SWR reading is at it's lowest point, below 1.5:1) - VE3AWA tested the antenna for me with his rig and right away he made contacts, one as far as West Virginia!

But me? Nothing, it's like I don't exist. It's so frustrating!

Take a power transformer with a center tapped high voltage winding. Build a push pull amp using the winding as a termination to  the modulator plates. 6146s or 6l6s will work. Now heising connect one of the modulator plates to the final. You will need a choke in the B+ lead.

I'm running out of parts now, I can't afford any hamfests until tax refund time next year.

So if anyone has an old broken or unwanted PA amp suitable for plate modulation (would prefer tube amp because transistor amps may rectify/become unstable near a transmitter I've found), or an old orphaned boat anchor of an AM transmitter their wife would be happy to see gone, I hope you keep me in mind.

I just ran a short QSO with VE3AWA today at 1pm EST on 7104. His noise level was S8, my carrier came in at S9 but voice quality was rough. I need to improve modulation somehow.

Thanks and 73s
geo

The nonlinearity in screen grid modulation occurs mostly near cutoff, so it should be OK on positive peaks until saturation resulting in flat-topping occurs.

Continental Electronics used a combination of control and screen grid modulation on some of their broadcast transmitters in the 50's and 60's.  This was supposed to be very linear, comparable to plate modulation. They always stayed away from high level plate modulation.

I recall Tom, K1JJ using a similar scheme  of control/screen grid modulation in the 1970's, with a 4-1000 IIRC.  Perhaps Tom could comment further on this.

I agree with Brian that most controlled carrier sounds like crap, although I have heard a few reasonably good sounding ones.  It works best with only 3 dB or so of carrier control, which still gives substantial saving in plate dissipation, safely allowing more output power on voice peaks, without generating the infamous pumping effect due to AGC action at the receiver.

Not all controlled carrier is screen grid modulation.  Before WW2, there were several widely used circuits that used plate modulation.  One scheme connected the DC plate circuit of the rf PA in series with that of the class B modulator stage.  As the modulator drew more current on voice syllables, its load impedance on the power supply decreased, correspondingly increasing modulator plate current and therefore rf plate current since both stages were connected in series across the power supply .  Another scheme used a saturable reactor that varied the a.c. voltage to the rf power supply plate transformer, controlled by the modulator plate current.

Has anyone ever modified the sreen grid modulation on a Knight T 150 would the new WC3K DX 60 modulation work on the 150.

John W9BFO

I tried Cathode modulation before, on the big 100+ watt breadboard rig, I ran into some problems with freq. drift and some internally generated noise/squealing caused by the modulator tube in some kind of reaction. I used a single 6AS7 for modulation in series with the 805's cathode, I used both sections of the 6AS7 in parallel.

I'm concerned that those problems might happen again if I try using Cathode modulation on this rig. Ideally I want to leave the Cathode/Control grid loop alone, isolated for RF purposes only - and apply modulation to some other part of the Xmitr.

I tried to see if I could make some kind of plate modulated thing happen this afternoon with a 35 watt solid state amp I have. But as soon as something RF starts near this amp, it becomes unstable and starts to break into HF oscillation on it's own, so that's a no-go there. I'm not even going to touch this small amp, it's just not worth the additional headache.

73s
geo

hi John ....just got your message ... just refreshed my memory with the WC3k mods to a dx 60

I think there may be a typo on the carrier control pot value ... 50 ohms seems way too low .... more like 5k or 50k

notice that the 6de7 stage a cathode is grounded .... this forms grid leak or contact bias which varies with audio level in and shifts plate voltage q point .... lift the cathode and place a 2k pot to ground and this will allow you to vary bias from grid leak to about 2 volts which should keep the plate q point steady and defeat the controlled carrier aspect .... some intermediate setting would allow some cont carrier op but not the usual 10 to 12 db these run which is annoying .... a single 6146 can gen about 15W carrier with plain screen grid mod ....hope this helps ...73 ... John

I found that when I tried the controlled carrier modulation, what was "modulation" was sort of a very abrupt analog to digital reaction, where my analog signal (audio) was being interpreted into a stream of highs and lows (as I saw on the scope). Sort of like a Pulse Width Modulation.

But the effect on the sound quality was absolutely horrible! It looked great on the plate current monitor but it was impossible to call it modulation. Digitization seems more accurate a description.

(http://www.noding.com/la8ak/images/2j.gif)

What I ended up trying was placing a 15k resistor in series with the first 6DE7's (the tube I ended up using was not a 6DE7, it was a 6BZ7/6BQ7, which is what I had on hand to try) cathode-ground connection and reduce the 22 Meg resistor down to something less than 1 meg. This let the first tube behave in a more linear fashion. But it was still not good enough to give me more than perhaps 40% modulation.

Somehow I'll have to do something to improve my situation. Just not sure what now.

73s
geo

Ding Ding Ding!! The light bulb goes off! Thank you John!  This certainly helps clear out some confusion in my mind!  Such a simple and elegant fix!

Geo, it is not my intention to hijack the thread you started, but I just felt that my questions were certainly on topic and relavant and the answers would help both of us!

73

John KX5JT

BFS John, for the SGM on the DX-60, are you speaking about the first 6DE7 (a) triode stage?  BTW, the pot on the cathode follower load in the WC3K mod should be a 50K ohm pot not a 50 ohm (replacing a fixed 33K ohm resistor).  A typo indeed.

TJL Terry.. I am intrigued by the feedback idea!  If you still own that modified DX-60 I'd love to make a schedule with you sometime to hear it on the air!

KYV Don... I recall you mentioning on a few occasions that you have increased the input impedance of your mic input stage to something on the order of 20 megs with a resistor to better match the Hi-Z D104 Crystal mic and get better low frequency response. Certainly this could apply here, but I wonder where on the input of the 12ax7 I would apply this? I'm assuming I would replace the 2.2 megohm from the grid to ground.

73 from JT John  

Don,
      you can easily get 100% in both directions if you bias the final well into class C, eliminate the grid leak resistor, and also modulate the control grid biass in phase with the screen. About 25% of the modulating voltage applied to the control grid works just fine.

I've been running my "scream modulated" 4-1000A like that for quite a few years now. It will go 100% negative without that funky crossover distortion normally found in screen mod rigs. It also has no problem going well over 100% positive as well..

there is a segment in the handbooks that says you should apply "a small amount" of the modulating voltage to the control grid as well.

I run 3200v on the plate, -150 or so on the control grid and around 290-300v on the screen. Drive it with a 6146 to 40 mills grid current, and load the hell out of it. It loafs along at a half gallon out and 100+% swingin the munky.

You and many others here have heard it, no one ever complains about my audio.

                                                         the Slab Bacon

Slab,

Now that you've settled on a good 4X1 screen modulated design - no grid leak, heavy class C, heavy loading, etc., what do you calculate the input/output final stage power % efficiency to be?  (Both dead carrier and peak power audio tone calculations)

T

Tom,
        I have never calculated the efficiency at peaks, but at resting carrier it is around
40-41% which is better than what the handbooks tell you it should be. IIRC the handbooks tell you you should be getting 30-33% efficiency. It is still loaded heavily enough to get well over 100% on peaks (on the scope). I think running the higher voltage has something to do with it.

Becuz the screen's impedance presents a varying load to the modder, I did wrap some feedback around it. I also installed 2 meter probe jacks on the output of the mod tranny to allow me to load down the modder as the plate voltage changes. (Higher Ep = lower Ig2) I just stab in a 10w dogbone to compensate for any necessary changes.

I have never ran any efficiency tests at peak audio. It of one of those things I want to do, bud never found that elusive little thing called a round tuit. It would be very interesting to know since this type of modulation is sometimes called "efficiency modulation."

However it has worked so well I have taken to "if it aint broke, dont fix it".

                                                            The slab bacon

Geo,
Do you need parts to build an AM rig?
What power level are you looking for?

In the US, most of the AM activity used to be around 7290, when the band was open, weekend mornings.
I have worked ve3owl? a number of times lately.

I have a bunch of parts for smaller type rigs I have no use for, share the wealth I say.
I could hook you up with tubes and mod trans if it helps.

Big stuff I dont got much of, but little parts I do.

Brett

here is a PDF of the basic design from it's basic inception done for me by Steve a few years back. I have made a few minor changes since this drawing, but this is basically it.

Thats a neat design!

Brett

the only difficult item to get is the ARC-5 mod tranny. But there are other ways around it. I used it cause I had it. that rig's been in operation close to 10 years now.

Yep, interesting design, Slab.

Back in the early 70's I had no mod iron. I once ran a 4X1 in class C with composite screen and grid modulation.  I used two standard (small) 8 ohm to 5K?  audio output transformers. These can be found in standard tube radios, etc.  I hooked the 5K sec of one xfmr to the 8 ohm primary of the other in parallel. Then fed this junction with a stanadrd SS PA amp (8 ohms output) and hooked the opposite 8 ohms in series with the grid and the other 5K in series with the screen.   I called it composite screen/grid modulation at the time. It worked FB.

To test your peak efficiency - it's easy.  Just do a sustained whistle to get a peak power reading on your peak reading meter and also note the PA current and voltage. Then do "peak power out /  (E X I) in ."   I expect it will be in the 60%-65% area. If it's more, then all the better. Your plate color will probably not change much since it's an efficiency technique.

Anyway, the reason I'm wondering is I find I can bias my linear into deep class C operation (without a grid leak) and on AM ONLY, get a nice linear waveform out approaching 65% eff - but the dead carrier eff seems better, like what you are seeing..  Being a triode there's no screen to modulate, so I'm wondering how close they are in overall similarity.  The linear's overall plate heat is reduced compared to running the linear in class B.  The constant carrier acts like a bias to idle the tube outa class C, so to speak. I usually add diodes to the fil CT to bias it heavily.

T

Really?
Seems odd!
But if it works, put a switch on it to remove the bias for AM use, and allow the diode bias for ssb use?

In my new 4x813 amp, I have ten 3 amp 1000 volt diodes in series to give 7.5 volts bias on the cathode.
( I got 100 of those diodes for $5.00 off ebay)

I get very little resting current without any drive, 0 to 40 ma, I never looked closely.

You are using an adjustable bias source and cranking it way up and get a clean signal out with modulation?

Brett

Brett,

I have about 200  3A 1KV diodes on a plexiglas board with a rotary switch to short some out as desired. They go in the fil CT as bias.  I can select from class A all the way down to deep class C.

Try it.... add lots of diodes until you are way below resting current into C on your linear amp. Then drive it with some carrier until you are at the normal carrier output level. You will need more drive than a normal class B config, of course. Then run some audio tones through. You will find they look FB.  This will not work on ssb, of course, and cause severe cross over distortion.

The AM carrier gives the final a "resting idle" to form the audio around on the load line. I believe the carrier is operating in class C, which helps the dead carrier eff, though the audio peaks are probably still in the class B eff category due to having to be linear.

Brett, please try it and take some careful measurements. If it works like I think mine does, then it will be worth publishing it around so that the many AMers using linears can increase their dead carrier efficiency. Right now, class B linear carrier eff at 25%-33% really sux... ;)

T

ok Geo ....the 6bz7/6bq7 is a small twin triode for low voltage / low power applications, not suitable for this job .... what you really need to have as the cathode follower that feeds the pa screen is a low mu triode or a tetrode connected as a low mu triode, quite similar to the discussions about driver tubes for class B modulators that have been here lately ... so what do you have ?  ....

I love when someone builds something, so I plan on sending geo a crap load of parts.
I have lots of small bits like mod transformers, tubes, sockets, meters, resistors, etc.

I will go through the junkbox and see what I have, I think I have a nice 120 watt multimatch mod trans, a bunch of old hifi type tubes I will never use, some dx100 parts, lots of 812 tubes, some 4D32 tubes, but no sockets for the 4d32's I think, a bunch of 6146's, plenty of 6as7's?, some 6b4's, and lots of other odds and ends type tubes.

I have a LOT of old meters.

So he could do plate modulation with something, lets say he uses a paair of 6146 tubes in the RF section, what to use as modulators?
I have some 807's I think, 6146's would work but tend to sound crappy, some little zero bias triodes and nix the screen and bias supplies?

Brett

I am plate modulating a Globe Chief (2 X 807) with an all transistor Radio Shack 40W PA amp I picked up at a garage sale for $3.   The mod transformer is from an old Johnson business band amplifier that used 2 X 2N3055 to modulate a 12V 6146.    I wonder if you could use an output transformer backwards with a solid state amp if it had adequate voltage and current rating.

I tried the circuit afterwards using a 6AS7. Not much difference in results.

Hi
Reading ,how you try to screen  modulate your transmitter,reading the ARRL hand book,it is the same as choke coupled modulator on the plate but used it at the screen .
in tha ARRL hand book it is called Clamp -tube modulation . but the choke is replaced by a resistor .

Just a though ,You used an 24 v 5 watt lamp as the Plate meter,is it possible that the Resistance of the lamp ,is high enough .So when the plate current of the 6146 gets high when You modulate the screen.There 's much drop in the Plate voltage of the 6146.
So it 's acting like a"negative feed back" on the RF output.

Gito.

It's a possibility but the lamp doesn't get bright/hot enough to illuminate to the 24 volt level it's designed to. So I don't think it's a major issue at this point.

Hi

Reading Your Article,modulating  the screen,it's the  right way,
so I"m still curios ,it is still possible that there is a" resistance" in your B+ or,it's the power supply that's goes" up and down" when you modulate your transmitter so the Effect is still " negative feed back" on Your RF output,when You modulate

Gito

I dug up a utc vm2 (new in box) a bunch of 807's and 1625's, some sort of utc s series driver transformer, and a bunch of other parts i have loads of, including some meters (ditch the bulb).

A lot to box up, so it might take me a while to ship it.

Brett

I took the boxes to fedex, and they wanted $230.00 to ship this stuff!
I then went to the post office, and it was not much cheaper.

I had no idea it cost so much to ship some junk, its not far from NJ...
Is there no way to get it to Canada without breaking the bank?

Brett

I had no idea it cost so much to ship some junk, its not far from NJ...
Is there no way to get it to Canada without breaking the bank?

Brett

Catapult?  ???

You've got a PM.

Anything over 150-200lbs it is definately MUCH MUCH cheaper to use truck freight.
they charge you by hundredweight. The initial minimum is kinda pricey, but after that it only costs a few more bux per 100lb after that.

Skip (K7YOO) sent me a "package down" here for Derb (and 1 piece for myself).
We had it shipped to my shop so I could unload / load it with a forklift.
The crate was somewhere over 400lbs and the shipping was somewhere around
$125 IIRC

                                                     The Slab Bacon

Well, I have 2 boxes, one is 40 pounds, the one with the iron in it, the other is under 20, (tubes and parts).
For that, $230.00!

Maybe because its going to Canada?

Brett

Scary!! Hmm, well then, don't worry about sending anything then, ok. I appreciate the offer very much, but these high prices today are killing everyone.

Hi

You have a 6146 transmitting tube and a 6L6 audio tube,here's a schematic diagram of Globe Scot 65 -plate modulated,Hopped it can be  used in your transmitter with a little change here and there,in the globe Scot 680 it used a 6AG7 as the oscillator.
 
Gito,

Gito - thanks! I'll see what I can do with it! I'm also using a 6AG7 osc. so it looks very hopeful! :)

73s
geo

Gito - the lower part of the schematic's parts list is missing, do you have the lower half of it?

Hi Geo

Looking at the schematic I Attach.it looks complete to me,but i"ll attached another one


Gito

Thanks, this one is better, full parts list, coil listing too!

73s
geo

Globe Scout 680 schematic diagram ,still used 6146 as final transmitter tube and 6L6 as modulator tube

the Glob Scrote makes a nice exciter for a leanyour amp.  The modulator circuit is full choke coupled Heising. Keep in mind, they only put out around 10-12w in phone mode.

Geo,

     Mr. Bacon correctly pointed out that this is Heising modulation, a form of plate modulation. I find it interesting that the schematics that Gito posted show NO screen modulation to the 6146 except for at the higher audio frequencies where a capacitive voltage divider from the plate to the screen is present. The screen dropping resistor to the 6146 (12K'ish) might be high enough so that the screen self modulates like a big Tetrode that has a screen choke present in series with the screen supply. I find the Globe Scout circuitry interesting.

    Keep in mind that Heising modulation without "tweaks" is incapable of more than about 80% modulation. The big limitation is the downward swing of the class A modulator tube cannot go to zero volts, and will likely not go much lower than 150-200 volts. Depending on your B+, your modulation percentage limit may not exceed 50%. All that said, a casual look at AM operators on the bands today, 50% modulation or lower is not uncommon. So you will have good company.  :P

    The modulator tube is running class A with cathode bias. A little negative feedback to a prior stage will straighten out a lot of the asymmetry that will be seen at the modulator output as you approach maximum modulation.

     The biggest obstacle I've experienced with commercial versions of Heising is that the Heising choke is undersized with an air gap to raise the saturation current point. That choke must bear the 6146 plate current (abt 100ma), and the modulator plate current (abt 50 ma). With a gapped-core choke at 8H @ 150ma, the reactance of the choke is pretty low at audio frequencies below 500 hz, so the audio modulator must restrict the low frequencies quite a bit. If you don't restrict the lows, then that 8H choke will load the 6L6 such that severe distortion will be observed. You could replace that choke with a 15 or 20H smoothing choke, and things would behave MUCH better. That choke would be pretty big though, and very heavy.

    I figured a way around this Heising choke size limitation with my Gonset G50 (search my posts here on AM FORUM) where you can use a center tapped choke (or the primary of a push pull transformer). Here the B+ goes to the CT, and one end goes to the modulator plate, and the other end goes to the RF stage modulated B+.  Now something magic happens. The RF PA current subtracts from the modulator current such that the choke magnetizing currents buck each other. In my case I use two 6L6's in parallel in the modulator that pull 100ma from the B+. The 6146 also pulls 100ma, and 100-100=0. This means that there is NO net magnetizing current in the choke. This also means that the choke does not need to be gapped, will have much higher inductance, will be much smaller and lighter. When I made this change I could for the first time modulate cleanly down to 50hz.

   My write up, circuits, pictures, scope traces, etc. are posted. Just search for them here.

Good Luck,
Jim
WD5JKO

Thanks! Sounds like a good approach to Heising!

Hi Geo

Yes it's Heising modulation,in the old books it's called Choke coupled modulator ,that says The audio output power of the modulator combined with the DC power in the plate circuit,is just as in the case of the transformer coupled modulator.how ever,there's considerably less freedom in adjustment,since no transformer is available for matching impedance(the transmitter tube and modulator tube).

 I attach the schematic drawing at it was.(Complete schematic diagram....covering the Globe Scout 65)

Yes Jim the choke must have large inductance, so when there's current  during positive going modulator excursions.the choke collapsing magnetic field pushes voltage above B+ rail.It is like a flyback transformer.
The Choke is the important part,it must have enough inductance,which must have the lowest flux leakage (no air gap).

In The ARRL handbook ,A complication is the fact that the plate voltage on the the modulator must be higher than the plate voltage on the RF amplifier,for 100% modulation.
There is an article  about a" modification  in the globe scout 65 in WRL word radio labratories constant current."
Just googling /searching in the internet  on   Heising Modulation.

In this article Heising modulation also known as Constant Current modulation.
In this article it used a resistor(450 Ohm in parallel  with a 10 ufd capasistor) drop the plate voltage of the RF amplifier(after the choke) /lower than plate voltage of the modulator tube ,moving the screen supply from the B+ to the plate voltage of the RF amplifier(that has been drop) ,so it's also modulated when the RF plate is modulated.
In this article the Author wrote ,he can get 18 to 20 watts carrier output in class C operation,and in the modified system allows 100% negative peak an nearly 100% positive peak

Gito

I made this 1 watt rig, Heising modulated, some months ago. I made no contacts of course with only 1 watt AM.

But what I found was that I could keep the plate voltages the same, only I had to make sure the modulator tube (far right) could sink at least double the current of the RF tube (bottom).

In this example, the RF tube can sink about 25mA and the Modulator tube can sink about 45mA. I can get up to and over 100% modulation on this tiny rig (but still make no contacts).

I'll keep it assembled if only as an experimental plaything.

Hi Geo

A nice looking transmitter you've made

Back to Heising modulation.
1. The ARRL books wrote to have the same impedance of the Mod tube and the impedance of the RF tube
2.It wrote the voltage developed by the modulator cannot swing to zero without a great deal of distortion.
3,Geo wrote he can get 100% with his 1 watt transmitter.
4.Jim used a center tap power amplifier (choke) to minimize magnetizing current.

Reading it ,in my opinion when Jim used a center tap Choke it's not Heising modulation anymore but plate modulation using a 1 t0 1 ratio modulation transformer.

So thinking of that, Heizing is a plate modulated transmitter using 1 to 1 ratio ,the same winding is used as the impedance of the RF tube and also the impedance of the Modulator tube (2 in 1)

Why it has to have the same Impedance  of the RF tube and Mod. Tube.

 As a normal plate modulated transmitter using a Modulator transformer ,the impedance of the modulator and the impedance of the Rf tube and Mod tube is very important.

A mod amplifier has a 3 k ohm impedance modulating a 6 k ohm RF amplifier has a winding turn is the root of 6:3 that is about 1,5 turn ratio.
So too fully modulate this transmitter  we must have the right turn ratio of the Modulator transformer,and also enough wattage Modulator.

In my opinion this is also implemented in Hezing modulation.
A class A modulator must have a certain impedance load,  Say it's 3 K ohm
and like above example the load impedance of the transmitter is 6 k ohm.

The modulator can not 100% mod the transmitter since it need a "step up" Transformer and in the "modulation transformer"(choke)  it's 1 to 1 ratio.
Imagine it as 2 different winding in parallel (actually it is only one winding).
So the Modulator can not supply the peak voltage that is needed for 100% modulation.
So we must first know the impedance of our Class A modulator and has enough audio power to modulate the transmitter
Than find a Transmitter tube loaded to the impedance value of the modulator tube.
So we have 1 to 1 impedance ratio and with enough audio power,hopefully we can drive 100% modulation .

Maybe I'm wrong

Gito

Gito,

  I like your analysis, and I think you are correct. With class A Heising, getting the largest undistorted swing out of the modulator is very important, and in most cases your limited to about 80% modulation.

  Some tweaks to Heising to get more modulation. Some can be stand alone or in combination:

1.) Drop the modulated B+ with a resistor, so the RF tube sees 10-20% less B+ than the modulator tube. Then bypass that resistor with a non polar cap so it is effectively bypassed for audio. The B+ is dropped, but the audio peak level is not dropped.

2.) Changed the modulator cathode bias to fixed bias. This puts the cathode at ground and increases the modulator plate to cathode voltage.

3.) Use a beam Power tube as the class A modulator tube. The plate can swing much lower than the screen voltage when compared to a pentode. A sweep tube like a 6DQ5 might be ideal here so long as we combine with item 3 below.

3.) As the plate swings low, compression becomes increasingly apparent, giving rise to 2nd HD. Use Negative Feedback to counter this.

4.) Use a low impedance audio driver so that the class A output can be run class A2 such that we can flow some G1 current to help get more plate swing in the downward direction.

5.) Return the modulator cathode to a low impedance - supply, such as -100 volts. This will increase the modulator plate to cathode voltage such that we can now swing more peak audio, and hit a higher modulation percentage. Getting the bias right will be more tricky though.

6.) Provide a sliding bias technique to shift the modulator bias point higher during high audio levels. Use this with items 3, 4 above.

7.) Use an oversized choke (20 H or more), or use the CT choke concept (oh, Gito says that is not Heising.. :))

Make any sense?
Jim
WD5JKO
I know, some readers might say, "switch the darn thing over to P-P plate modulation". Yea but that is no fun. ::)

Regards,
Jim
WD5JKO

 Hi Jim

Your analysis makes a lot of sense to me.
But Jim ,not all of Us has the Skill and knowledge like You  to Implant it on Our Transmitter.
I agree with all that You have wrote.
But one thing changing the bias ,does not it change the operating class? and changing the Impedance of Modulator Tube?Because in Heising "actually" is Plate modulation with a "modulator transformer" with 1 to 1 Ratio

Regards

Gito

This is an old thread but I was looking at some of the theories put forth here on screen modulation.

Recently in Electric Radio there was an article on a dual final CW rig using 4D32s.

Sooo, I though maybe something like this might work, IE taking the ER article and modulating the screen grids.

This is only on paper so far and I haven't implement it. Some tweeking of component values may have to done.

In June 2014 issue Page 6, the screen grid voltage is regulated by a high voltage bipolar transistor.

What this circuit does is further solid state some circuits in order to modulate the screen grid.

NMOS M2 Modulates the voltage at M3's gate. M3's source provides the varying audio voltage to the screens. R13 limits the DC screen current to about 70mA max.

R8 sets the bias on M3's gate, and via DC and AC feedback from the source, tends to stabilize the bias and reduce distortion.

M1 replaces the 2E26 for screen grid protection in case of RF drive loss.

D23 insures he screen voltage never tops 350 volts.


Anyway, something to think about for next winter's projects.

Phil - AC0OB

Phil

High voltage FETs often have a very high gate-to-source capacitance (around 3000pF), and therefore a fairly low gate-to-source impedance at audio frequencies of a few kHz.

The source impedance of the circuit driving the gate of M3 (47k ohms) may be too high to produce a flat screen modulation v. frequency response.

Stu

   It appears to me that m3 is in a source follower or common drain configuration. Wouldn't the ciss input capacitance of that FET be negated as the gain of the fet approaches 1?

   Looking at the circuit makes me want to build it. I wonder if D1 or D3 might clip on audio peaks depending on audio drive level and pot R8 setting?

Good Stuff.
Jim
Wd5JKO

Jim

You are correct.

Since the AC part of the gate-to-source voltage will be only a tiny fraction of the AC part of the gate-to-ground voltage... the gate-to-source capacitance will not load down (i.e. draw excessive current from) the circuit that is driving the gate of the source follower.

Thanks for pointing this out.

Stu

The M3 fet is a source follower configuration. We are not trying to saturate it with a fast rising square wave, but rather allow it to operate in it's linear region with a relatively slow rising and falling complex audio wave. For fet gates, this voltage represents a voltage range of approx. 1.0V to 4V gate-to-source voltage. I have tested these high voltage fets and found that to be the case.

D1 keeps the M2 fet from saturating, and D2,3 keeps the screen voltage from rising to severe levels.

I would expect the bias to be set so the quiescent voltage level on the screen is about 250 volts or thereabouts. Of course, R8 also sets the power level.

I think I will add an audio level control to allow the input audio level to be attenuated.

Again, a work in progress, but I appreciate the comments.

Phil - AC0OB